From what I gather, they use the open source Ceph project to manage their file system. There was an update to Ceph that they applied and that promptly corrupted the shiz out of all the files on their servers resulting in literally every service they offer horribly breaking... It really sounds like every IT persons worst nightmare come true.
With the shear amount of data they have on their servers I imagine it might be another week before they get everything sorted out.
edit: Also looking at their list of things that are working it looks like things that would have smaller overall data are being restored first. So I imagine they have automated scripts running 24/7 and whatever gets done first goes back up.